Video Editing Workflow
Revise or restyle an existing video clip on the Astorie canvas with video to video AI. Connect a source clip to a Video node in Video Editing mode, describe a focused visual change or supply a style reference and compare the generated take frame by frame against your original footage.
Video-to-video generation starts from motion, timing and framing that already exist in your source clip. Comparing a broad, unguided restyle with a targeted canvas edit shows how keeping the original footage nearby helps you preserve the movement that matters.
Start with a clean source clip that you have permission to edit and a clear idea of the single visual change you want to test.
Select a short, well-lit video clip with readable subject motion and clean framing. Keep the untouched original on the Astorie canvas so you can compare timing and details after generation.
Add a Video node and connect your source video to its input port. In the documented Video node workflow, a Video connection sets the node to Video Editing mode. Add a reference image or text brief if your chosen model supports guided restyling.
Choose a video model on the canvas that supports video-to-video editing, such as Runway Aleph where available in your workspace, and run the generation step.
Play the generated take and the source clip side by side. Check that the core movement, subject outlines and frame timing hold up before exporting the revised shot for your sequence.
Inspect three areas across the full duration of the clip before replacing the original shot in your cut.
Watch key gestures, camera pans and contact points alongside the source video. Confirm that the edited version follows the original rhythm rather than drifting out of sync mid-shot.
Look closely at hair, hands, clothing edges and fast-moving props. Video-to-video transformations can introduce shimmer or edge crawl where the new visual style meets a complex background.
Check that the new surface look or lighting direction stays steady from the opening frame to the end of the clip rather than shifting brightness or hue as the camera moves.
Keep your source footage, model selection and quality checks aligned across the canvas workflow.
| Input | A permitted source video clip connected to a Video node in Video Editing mode, plus a focused edit prompt or supported reference image. | |
|---|---|---|
| Output | A transformed or restyled video clip on the canvas for frame-by-frame comparison with your original footage. | |
| Review | Compare motion timing, subject identity, edge stability and lighting consistency against the untouched source clip. | |
| Plans and credits | Video-to-video runs consume workspace credits based on the selected model and clip settings; check the pricing page for plan details. | |
| Rights and finishing | Use only source footage you have rights to modify, and handle final trimming, audio sync and color finishing in your NLE. |
Explore documented video editing models and connected canvas tools on Astorie.
See how Runway Aleph edits and reskins existing footage on the canvas while working from source motion and reference images.
Review how connecting a Video input triggers Video Editing mode and how other node connections work.
Prepare approved video-to-video takes and original source references for your nonlinear editor.
Text-to-video invents a clip from written words, and image-to-video animates a still frame from scratch. Video to video AI starts from an existing moving clip connected to a Video node so you can transform its visual treatment or scene details while building on the source motion and framing.
No. Video-to-video editing requires a model and node mode that accept a video input, such as the Video Editing mode described in the Video node documentation and in-context editing models like Runway Aleph. Check the supported inputs of your selected model on the canvas before running a shot.
No model guarantees pixel-exact preservation across an entire moving shot. Treat each video-to-video run as a creative transformation, keep your untouched original clip nearby and inspect important faces, products and edges frame by frame.
When you change lighting, costume, background and camera framing in a single request, it is hard to tell which part of the prompt caused flicker or drift. Testing one focused change per run keeps the source motion clearer and makes revisions much easier to evaluate.
Bring a source clip onto the Astorie canvas, connect a Video node in Video Editing mode and compare the transformed take against your original footage.
